The results
81%
reduction in customer waiting time
91 s
per automated investigation, down from 8 minutes
67%
of low-risk alerts closed automatically
50%
shorter queues at retail branches
Key takeaways
- A global foreign exchange leader deployed Ray, ThetaRay's agentic AI investigation layer, to handle a growing backlog of low-risk compliance alerts.
- Automated investigations now take 91 seconds instead of 8 minutes, an 81% reduction in the time customers wait for a compliance decision.
- Ray unified ThetaRay detection, Casmex, NICE Actimize and internal screening tools under a single agentic investigation layer without ripping and replacing existing infrastructure.
- 2/3 of low-risk alerts (a 67% policy-approved closure rate) are now investigated and closed by Ray within the policy the compliance team approved, and branch queues are 50% shorter.
- The engagement moved from kickoff at the end of March to go-live in May.
Executive summary
Rapid global expansion and soaring transaction volumes often push legacy anti-money laundering (AML) and compliance frameworks to their breaking point. For one of the world's largest independent foreign exchange and financial services brands, an explosion of low-risk compliance alerts began outstripping human capacity. Fragmented across multiple legacy vendors, investigations required tedious, multi-system manual deep-dives, creating a massive backlog and bottlenecking retail operations.
By deploying Ray, ThetaRay's agentic AI investigation layer, the institution unified its fragmented compliance architecture (NICE Actimize, Casmex and internal screening tools) under a single agentic investigation layer.
The impact was immediate: automated investigations dropped from 8 minutes to 91 seconds, cutting customer waiting time by 81%. 2/3 of low-risk alerts are now investigated and closed within approved policy, driving a direct uplift in foreign exchange (FX) and visa revenue.
8 min
per investigation before Ray
91 s
per automated investigation with Ray
81%
less time customers wait
The challenge: a fragmented architecture buckling under volume
Operating across high-traffic global hubs and international airport terminals, the institution faced immense pressure to maintain stringent compliance standards without compromising the customer experience. However, its existing compliance stack created severe operational friction:
- Slow, manual investigation paths: Compliance analysts spent 5 to 30 minutes per alert manually gathering data and cross-referencing files.
- The silo problem: Fragmented investigations required analysts to log into multiple systems across different vendors, slowing down decision-making.
- Volume and queue pressure: An expanding backlog of alerts grew faster than compliance headcount could scale. In the real world, this backend bottleneck showed up as peak-hour customer queues at brick-and-mortar branches and airport terminals.
The solution: unifying the stack with agentic AI
To solve this, the institution deployed Ray as a centralized, cross-platform orchestration layer.
Unlike traditional upgrades that require ripping and replacing existing infrastructure, Ray overlays the company's mixed-vendor ecosystem, unifying ThetaRay detection, Casmex and NICE Actimize into a single workflow.
Key capabilities deployed
- End-to-end autonomous investigation: For low-risk anomalies, Ray executes the entire investigation workflow (gathering data, analyzing context and providing clear decisioning) in about 91 seconds, leaving zero humans in the loop for low-risk clears.
- Cross-platform architecture: Instead of navigating disparate tools, compliance analysts use a single interface where Ray aggregates risk indicators and visualizes transaction data paths across all legacy platforms (ThetaRay, Casmex and NICE Actimize) simultaneously.
Business impact: measurable operational and commercial ROI
Within months of kickoff and workshop alignment, the institution transformed its compliance department from an operational bottleneck into a commercial enabler:
- 81% less waiting time: Automated investigations went from 8 minutes to 91 seconds, so customers get a compliance decision in about a minute and a half.
- 67% policy-approved closure rate: 2/3 of all low-risk alerts are now investigated and closed by Ray within approved policy, eliminating alert fatigue.
- 50% queue reduction: Streamlining the backend compliance process cut queues at physical retail branches in half.
- Top-line revenue growth: Faster processing times unlocked a noticeable uplift in high-margin foreign exchange (FX) transactions and temporary visa processing volumes.
“AI didn't replace our analysts. It freed them to do the job they were actually hired to do, focusing human intellect on truly complex, high-risk financial crime while the platform handles the noise.”
VP of Compliance & MLRO, Global FX
How long did the Ray deployment take?
The transformation from kickoff to deployment was realized over a rapid, phased timeline:
- Phase 1 (end of March): Project kickoff and scope definition.
- Phase 2 (mid-April): Core alignment and multi-vendor system mapping meetings.
- Phase 3 (late April): Intensive technical workshops to configure Ray's agentic playbooks.
- Phase 4 (May): Official system "drop" and go-live deployment.
A question for compliance leaders
When your compliance team is forced to hop between a legacy case manager, screening lists and internal databases, how many hours of productive investigation time are lost to manual data gathering each week? And is that backend friction secretly driving away your customers at the point of sale?
Frequently asked questions
How much did Ray reduce customer waiting times for the foreign exchange leader?
Ray cut automated investigations at the foreign exchange leader from 8 minutes to 91 seconds, an 81% reduction in the time customers wait for a compliance decision.
What is agentic AI investigation for AML alerts?
For the foreign exchange leader, Ray, ThetaRay's agentic AI investigation layer, executes the entire investigation workflow for low-risk anomalies: gathering data, analyzing context and providing clear decisioning, with zero humans in the loop for low-risk clears.
Can Ray work with existing AML vendors like NICE Actimize?
Yes. Ray overlays the institution's mixed-vendor ecosystem without ripping and replacing infrastructure, unifying ThetaRay detection, Casmex and NICE Actimize into a single workflow.
What share of alerts does Ray close automatically for the foreign exchange leader?
Ray achieved a 67% policy-approved closure rate at the foreign exchange leader, meaning 2/3 of all low-risk alerts are now investigated and closed by Ray within the policy the compliance team approved.
How long did the Ray implementation take?
The Ray implementation moved from project kickoff at the end of March, through system mapping in mid-April and playbook workshops in late April, to go-live in May.
